Qvest – Software-defined production in practice
Patrick Steinert, Head of Technology Consulting, Qvest
The Dynamic Media Facility provides a blueprint for moving production from dedicated infrastructure to shared, software-defined platforms. A multi-vendor reference implementation now puts that blueprint to work and shows what changes when architecture meets day-to-day production.
From concept to production
For years, software-defined production has been discussed largely in terms of architecture. That is changing. Broadcasters and media organizations now need to understand how these principles perform against the realities of production.
The pressure is familiar: more output, more formats, shorter launch cycles, and tight budgets. Live events add major peaks in demand, while hybrid operations increase the number of technologies that need to work together. Fixed production stacks, often sized for exceptional workloads, are increasingly difficult to align with these requirements.
Software-defined production offers another approach. Production capabilities can be configured around current demand instead of remaining permanently tied to dedicated infrastructure. The important test is whether this works reliably across real workflows, technologies, and operating models.
What Software-Defined Production Changes
The EBU’s Dynamic Media Facility (DMF) provides a framework for this model. Capabilities such as mixing, replay, graphics, clipping, transcoding, monitoring, and playout can run as software-based media functions on shared infrastructure rather than on dedicated systems.
The Media eXchange Layer (MXL) provides the open exchange mechanism between those functions, allowing video, audio, and timed metadata to move efficiently within the platform and supporting interoperability between applications from different vendors.
As a result, production functions are no longer permanently bound to a particular device, room, or workflow stack. They can be deployed, started, stopped, monitored, and released according to production needs. Resources are shared, while workflow configurations can be adapted and reused.
Existing technologies remain part of this picture. SDI and ST 2110 continue to serve important roles in ingest, contribution, and egress. A DMF-based approach allows media organizations to introduce new capabilities alongside existing investments across on-premises, edge, cloud, and hybrid environments.
One Platform, Multiple Productions
A scheduled studio production, a news operation, and a major sports event have little in common when it comes to workload. Studio production is predictable. News needs to respond quickly to changing sources and events. Sports creates short periods of intense demand with replay, graphics, highlights, monitoring, and multiple outputs.
Traditionally, these requirements often lead to separate technical stacks, each with its own capacity planning. The Qvest DMF Showcase demonstrates another model. The same environment can switch between vodcast, news, and sports workflows, with the required media functions configured for each scenario.
That changes how capacity is used. Resources do not have to remain permanently assigned to one type of production. Proven workflow patterns can be reused, temporary setups can be prepared faster, and capacity that is not needed by one production can support another.
The benefit is straightforward: infrastructure follows the production requirement instead of the production being constrained by the infrastructure available to it.
Interoperability is the Real Test
Flexibility depends on interoperability. Moving production from proprietary hardware into a closed software ecosystem would solve one dependency only to create another.
For that reason, the DMF Showcase is built as a multi-vendor implementation. It combines orchestration and lifecycle management, live video processing, audio processing, denoising, monitoring and multiviewing, and networking in one shared setup.
MXL provides the common media exchange foundation. Qvest’s open-source mxl-k8s extension enables scalable use of MXL across multiple nodes in a cluster, while go-mxl simplifies integration with modern software stacks.
What matters is how the components behave together. The DMF Showcase allows signal flows, orchestration, monitoring, lifecycle management, and interoperability to be tested within the same setup. Integration issues can be identified early, and technology providers can see how their components operate as part of a wider architecture.
Standards create the common foundation. Implementations like this provide the experience needed to turn that foundation into production systems.
Operating software-defined production
A flexible technical architecture still needs disciplined operations. Dynamically provisioned media functions require monitoring, identity and access management, security, cost transparency, support processes, and clear ownership.
Production scheduling and resource orchestration also become more closely connected. Broadcast engineering, IT, cloud, software, DevOps, production operations, and vendor management increasingly work across the same platform. Teams need visibility into how infrastructure, applications, and workflows behave together, particularly when resources are allocated dynamically.
This makes controlled adoption important. A focused use case can first validate the architecture and interoperability. A pilot can then test the workflow and operating model under realistic conditions before moving toward a first production or rehearsal.
Trying to prove every possible use case at the outset adds complexity without necessarily adding confidence. More useful is a sequence of clearly defined steps, with lessons from each stage carried into the next.
From Implementation to Adoption
The development of software-defined production depends on more than technology from individual vendors. Broadcasters, technology providers, integrators, and industry bodies need to test common approaches together.
Reference implementations provide a place to do that. They show which parts of the architecture are already working well, where integration still requires effort, and which operating questions need to be addressed before broader deployment.
Qvest contributes to this work through the DMF Showcase, participation in the AMWA-EBU Joint Task Force on Dynamic Media Facilities, and open-source contributions to the Linux Foundation MXL Project.
The industry now has an opportunity to move the discussion forward: from defining software-defined production to learning where it creates measurable value, how open interoperability can be maintained, and what is required to operate it reliably at scale.


























